3
Dispersant-Oil Interactions and Effectiveness Testing

Dispersants are mixtures of solvents, surfactants, and other additives that are applied to oil slicks to reduce the oil-water interfacial tension (NRC, 1989; Clayton et al., 1993). Interfacial tension is the free energy change that is associated with a change in the contact area at the interface between two immiscible phases (e.g., solid-liquid, liquid-liquid, liquid-gas). The term surface tension is also used to describe this phenomenon. Although these two terms are often used interchangeably, interfacial tension is considered to be the more general term, which can be applied to describe the free energy at the interface between any two phases, whereas surface tension applies specifically to those cases in which one of the phases is a gas (Lyklema, 2000). Reduction of the interfacial tension between oil and water by addition of a dispersant promotes the formation of a larger number of small oil droplets when surface waves entrain oil into the water column. These small submerged oil droplets are then subject to transport by subsurface currents and other natural removal processes, such as dissolution, volatilization from the water surface, biodegradation, and sedimentation resulting from interactions with suspended particulate material (SPM).

For the purpose of this and subsequent discussions, it is important to define two terms that are used interchangeably in the dispersant literature: entrainment and dispersion. In this report, entrainment is specifically the transport of oil from a surface slick into the water column by wind and waves, while dispersion includes both entrainment and subsurface transport (mixing and advection) by turbulent forces. It should also be mentioned that in the hydrodynamics literature the term dispersion



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects 3 Dispersant-Oil Interactions and Effectiveness Testing Dispersants are mixtures of solvents, surfactants, and other additives that are applied to oil slicks to reduce the oil-water interfacial tension (NRC, 1989; Clayton et al., 1993). Interfacial tension is the free energy change that is associated with a change in the contact area at the interface between two immiscible phases (e.g., solid-liquid, liquid-liquid, liquid-gas). The term surface tension is also used to describe this phenomenon. Although these two terms are often used interchangeably, interfacial tension is considered to be the more general term, which can be applied to describe the free energy at the interface between any two phases, whereas surface tension applies specifically to those cases in which one of the phases is a gas (Lyklema, 2000). Reduction of the interfacial tension between oil and water by addition of a dispersant promotes the formation of a larger number of small oil droplets when surface waves entrain oil into the water column. These small submerged oil droplets are then subject to transport by subsurface currents and other natural removal processes, such as dissolution, volatilization from the water surface, biodegradation, and sedimentation resulting from interactions with suspended particulate material (SPM). For the purpose of this and subsequent discussions, it is important to define two terms that are used interchangeably in the dispersant literature: entrainment and dispersion. In this report, entrainment is specifically the transport of oil from a surface slick into the water column by wind and waves, while dispersion includes both entrainment and subsurface transport (mixing and advection) by turbulent forces. It should also be mentioned that in the hydrodynamics literature the term dispersion

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects (sometimes shear dispersion) refers to a specific mixing process resulting from the combination of shear in the mean velocity coupled with turbulent mixing (or other transport mechanism) in the direction of the shear. This process will be discussed in Chapter 4 and will be denoted as hydrodynamic dispersion to avoid confusion. The following sections address dispersant chemistry, the physical and chemical interactions of dispersants with oil slicks and droplets, oil chemistry and weathering behavior and how they affect the window of opportunity for effective dispersant applications, and the importance of turbulence for introducing the energy necessary to entrain oil droplets into the water column as well as their subsequent transport by dispersive and advective processes. Next is a discussion of effectiveness testing and related issues, including laboratory systems, wave-tank tests, field studies, and studies involving spills of opportunity. Several of these topics are only considered briefly because there are a number of excellent reviews that consider the mechanisms of dispersant action and laboratory and field testing of dispersant performance (e.g., Meeks, 1981; Rewick et al., 1981; Mackay et al., 1984; Nichols and Parker, 1985; NRC, 1989; Clayton et al., 1993; Trudel, 1998; Etkin, 1999). Topics for which there are still major uncertainties or where data gaps exist are considered in greater detail, along with explicit findings and recommendations for areas requiring additional research. COMMERCIAL DISPERSANT PRODUCTS AVAILABLE FOR USE IN U.S. WATERS A typical commercial dispersant is a mixture of three types of chemicals: solvents, additives, and most importantly, surface-active agents (i.e., surfactants). Solvents are added primarily to promote the dissolution of surfactants and additives into a homogeneous dispersant mixture. In addition to keeping the surfactants in solution, these solvents reduce the product’s viscosity and affect the dispersant’s solubility in oil. Also, solvents determine to what extent the dispersant may be premixed with water for some spraying applications. Because aqueous-based solvent systems freeze in spray nozzles at ambient temperatures below 0° C (roughly 32° F) their usefulness is often limited in arctic or subarctic environments. Additives may be present for a number of purposes, such as improving the dissolution of the surfactants into an oil slick and increasing the long-term stability of the dispersant formulation. Surfactants are compounds containing both oil-compatible (i.e., lipophilic or hydrophobic) and water-compatible (i.e., hydrophilic) groups. Because of this amphiphatic nature (i.e., opposing solubility tendencies), the surfactant molecules will reside at the oil-water interface as shown in

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects Figure 3-1. The surfactant reduces the oil-water interfacial tension by orienting with the hydrophilic groups interacting with the water phase and the hydrophobic groups interacting with the oil. Reduction of the oil-water interfacial tension facilitates the formation of a large number of small oil droplets that can be entrained into the water column. Commercial formulations of modern chemical dispersants are usually comprised of two or more surfactant molecules that have differing solubilities in both water and oil. One parameter that has been used to characterize these different solubilities is the hydrophile-lipophile balance (HLB). The HLB ranges from 0 (no hydrophilic group) to 20 (no hydrophobic group), and the specific value characterizes the tendency of the surfactant to preferentially dissolve in either the oil phase (low HLB) or the aqueous phase (high HLB). The dominant group of the surfactant molecule will tend to orient in the outer phase to form a droplet of either oil or water (Porter, 1991). Therefore, a predominantly lipophilic surfactant (with a HLB below 7) will favor water-in-oil emulsions (mousse) where oil forms the continuous phase with discrete water droplets entrained within it (Porter, 1991). Natural components that promote the for- FIGURE 3-1 Mechanism of chemical dispersion: surfactant accumulates at oil-water interface, facilitating formation of small oil droplets that become entrained in the water column. Blow-up of oil droplet shows orientation of surfactant at the droplet surface with the hydrophilic group projecting into the water phase and the lipophilic group projecting into the oil phase.

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects mation of mousse (e.g., the resin and asphaltene fractions of crude oil) are generally lipophilic. In contrast, a predominantly hydrophilic surfactant (with an HLB greater than 7) will favor oil-in-water dispersions (i.e., entrained oil droplets in a water body) (Porter, 1991). The blend of surfactants in commercial dispersant formulations tend to be hydrophilic and the current formulations usually consist of surfactant mixtures with an overall HLB in the range of 9 to 11 (Clayton et al., 1993). An example of the orientation of surfactant molecules at the oil-water interface is presented in Figure 3-2. Compound A is sorbitan monooleate (HLB = 4.3; predominantly lipophilic). Compound B is similar to A but has been ethoxylated with molecules of ethylene oxide to make it more hydrophilic (HLB = 15). The dispersant formulation shown in Figure 3-2 contains more compound B than A. Such a balance will promote formation of stable oil-in-water dispersions (entrained oil droplets in the water column) because the dominant hydrophilic group of the surfactant mixture favors the formation of oil droplets in water. The use of two or more surfactants with differing HLB values, but an overall average HLB in the range of 9-11, allows for closer physical interactions (i.e., packing) of the surfactant molecules at the oil-water interface compared to a single surfactant with an HLB value in this range (Porter, 1991). This produces a stronger interfacial surfactant film. Although ionic surfactants can inhibit coalescence of small droplets into larger droplets that would resurface more quickly by providing an electrostatic repulsion barrier (Porter, 1991), recent measurements suggest that this barrier is too small to significantly affect the collision efficiency (i.e., the fraction of collisions that result in coalescence), at least for dispersants (e.g., Corexit 9500) that consist mainly of nonionic surfactants, even when the dispersant-to-oil ratio (1:10) is relatively high (Sterling et al., 2004c). Exact compositions for commercial dispersant formulations are proprietary, but their generic chemical characteristics are broadly known (e.g., Wells et al., 1985; Brochu, et al., 1986; NRC, 1989; Fingas et al., 1990; Singer et al., 1991, 1996; George-Ares and Clark, 2000). In general, a limited number of surfactant agents are currently used. Current dispersant formulations consist of mixtures of one or more surfactants, which may be either nonionic or anionic. Cationic (positively charged) surfactants are not used in current formulations (Clayton et al., 1993) because they are usually quaternary ammonium salts that are inherently toxic to many organisms. The Corexit products are by far the most prevalent of all dispersants held in industry stockpiles within the United States, making up as much as 95 percent is some instances (J. Clark, ExxonMobil Research and Engineering Company, Fairfax, Virginia, written communication, 2005). Corexit 9527 was developed in the 1980s; it was supplemented in the 1990s

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects FIGURE 3-2 Orientation of surfactants at oil-water interface in dispersed oil droplets. Surfactant A is sorbitan monooleate (a.k.a., Span 80; HLB ≈ 4.3); surfactant B is ethoxylated (E20) sorbitan monooleate (a.k.a., Tween 80; HLB ≈ 15). by the introduction of Corexit 9500, which includes the same surfactants incorporated into a different solvent (George-Ares and Clark, 2000). Both products contain a mixture of nonionic (48 percent) and anionic (35 percent) surfactants. The major nonionic surfactants include ethoxylated sorbitan mono- and trioleates and sorbitan monooleate; the major ionic sur-

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects factant is sodium dioctyl sulfosuccinate (Singer et al., 1991). Neither Corexit product contains polyethoxylated alkyl phenols (J. Clark, ExxonMobil Research and Engineering Company, Fairfax, Virginia, written communication, 2004). A different solvent was used in Corexit 9500 for two reasons. First, prolonged exposure to Corexit 9527 caused adverse health effects in some responders. These effects were attributed to its glycol ether solvent (2-butoxyethanol). Therefore, the solvent was replaced by a mixture of food-grade aliphatic hydrocarbons (Norpar 13; n-alkanes ranging from nonane to hexadecane) in Corexit 9500 (Varadaraj et al., 1995). The second reason for changing the solvent in the reformulated dispersant was to extend the window of opportunity for dispersant use. This window of opportunity is limited by the effects of weathering on the chemical and physical properties of the spilled oil, especially the increase in oil viscosity. Corexit 9500 has been shown to be slightly more effective with high-viscosity oils than Corexit 9527. THE PHYSICAL CHEMISTRY OF DISPERSANT-OIL INTERACTIONS AND THE ENERGY REQUIREMENTS FOR EFFECTIVE OIL-DROPLET ENTRAINMENT AND DISPERSION The objective of an oil-spill dispersant application is to lower the oil/water interfacial tension to enhance entrainment of small oil droplets into the water column at lower energy inputs. Entrainment of small oil droplets into the water column (by either physical or chemical means) increases the oil-water interfacial area, which as shown in Eq. (3-1), requires energy: (3-1) where WK is the mixing energy (ergs or g-cm2-s−2; 1 erg equals 10−7 joule (kg-m2-s−2)), γo/w is the oil-water interfacial tension (dynes-cm−1, where 1 dyne equals 1 g-cm-s−2; equivalent to ergs-cm−2), and Ao/w is the oil-water interfacial area (cm2). Therefore, reduction of the oil-water interfacial tension allows creation of a larger amount of interfacial area for the same level of energy input. Note that Eq. (3-1) provides an estimate of the minimum energy input that is required to disperse oil as droplets in the water column. Additional energy, which is proportional to viscosity, will be required to form droplets by stretching a continuous oil layer to the point at which it breaks. The seven requirements for a chemical dispersant to enhance the formation of oil droplets (NRC, 1989) are: The dispersant must hit the target oil at the desired dosage.

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects The surfactant molecules in the dispersant must have time to penetrate and mix into the oil. The surfactant molecules must orient at the oil-water interface with the hydrophilic groups in the water phase and the lipophilic groups in the oil phase. The oil-water interfacial tension must decrease due to the presence of the surfactant molecules at the oil-water interface, thereby weakening the cohesive strength of the oil film. Sufficient mixing energy must be applied at the oil-water interface (by wind and/or wave action) to allow generation of smaller oil droplets (with a concomitant increase in interfacial surface area). The droplets must be dispersed throughout the water column by a combination of diffusive and advective processes to minimize droplet-droplet collisions and coalescence to form larger droplets (which can resurface in the absence of continued turbulence). After entrainment, the droplets must be diluted to nontoxic concentrations and remain suspended in the water column long enough for the majority of the oil to be biodegraded. Turbulent energy is the environmental parameter most responsible for generating and transporting dispersed oil droplets in the ocean. Delvigne and Sweeney (1988) studied natural dispersion and argue that the smallest scales of turbulence, with the greatest shear, are responsible for initial droplet formation, while the larger eddy scales are responsible for the subsequent vertical transport (described in more detail in Chapter 4—Transport and Fate). Conversely, Li and Garrett (1998) argue that natural dispersion is generated mainly by dynamic pressures associated with larger eddy scales, resulting in the creation of relatively large droplets (i.e., order of 100 µm diameter) that resurface relatively quickly. They suggest that reduction of the oil-water interfacial tension by chemical dispersants allows the mechanism of turbulent shear to govern droplet formation, which leads to smaller droplets (i.e., order of 10 µm diameter), which is more consistent with the diameters observed for “permanently dispersed” droplets. Unfortunately, the droplet-size distributions of chemically dispersed oil have only rarely been compared directly to those produced when untreated oil was dispersed under identical conditions (see Box 3-1). In the few cases where direct comparisons were made, however, the volume mean diameter was reduced by 30–40 percent by dispersants (Jasper et al., 1978; Lunel, 1995b). Figure 3-3, which was reconstructed from data presented by Lunel (1995b), shows the effect of a chemical dispersant (premixed Dasic Slickgone NS) on the droplet-size distribution produced when Forties crude oil was dispersed at sea: the number of small droplets (<50 mm) increased by about 5- to 30-fold,

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects BOX 3-1 Droplet-Size Distributions: What Are They and Why Are They Important? When oil is entrained in the water column due to input of turbulent energy, droplets of various sizes are produced, regardless of whether the process is enhanced by addition of dispersants. Droplet-size distributions describe the relative abundance of droplets of various sizes, which may range from <1 µm to >100 µm in diameter. These distributions can be based on either droplet number or volume, although the volume distribution may be most informative, because the relationship between droplet volume and oil mass is constant regardless of droplet size (i.e., the proportionality constant is the density), whereas the relationship between droplet number and oil mass is not. The most common metrics for characterizing the central tendency of droplet-size distributions are the mean and median diameter, which will be approximately the same if the droplet sizes are normally distributed. The number mean diameter (NMD) is a simple average of droplet diameters, whereas the volume mean diameter (VMD) is the diameter of a droplet with the average volume (i.e., the mean of the volume distribution): (3-2) where ni is the number of droplets with diameter Di. The VMD is larger than the NMD. Number and volume median diameters (also commonly referred to as NMD and VMD) are those droplet diameters that divide the number and volume distributions in half (i.e., 50 percent of the oil volume is present as droplets smaller than the volume median diameter). whereas the number of large droplets (>50 mm) produced from dispersant-treated and untreated oil were similar. Note that although there were relatively few very large droplets produced from either treatment, these represented a significant fraction of the oil mass in both treatments, because the volume of oil in each droplet is proportional to the diameter cubed. Therefore, the volume distribution is extremely sensitive to uncertainty in the number of large droplets. This uncertainty can be seen in the reconstructed volume distribution shown in Figure 3-3. It is not clear whether the differences in characteristic droplet size are statistically significant, but if real, they would result in a 50–65 percent decrease in droplet rise velocity. Therefore, this phenomenon is potentially important and should be investigated further.

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects Droplet-size distributions result from the interaction of two processes: (1) droplet formation due to turbulent shear and (2) size fractionation due to differential rise velocities (Lunel, 1995b). Although the mechanism of droplet formation has not been proven, the initial size distribution of chemically dispersed oil droplets is thought to be related to the scale of the smallest eddies (i.e., microscale turbulence; Delvigne and Sweeney, 1988; Lunel, 1995b; Li and Garrett, 1998), but the distribution will be shifted toward smaller droplets following a period of quiescence due to resurfacing of larger droplets (Daling et al., 1990; Lunel, 1995b). Lunel (1995b) has suggested that dispersant effectiveness tests should be conducted in laboratory-scale systems and wave tanks that generate microscale turbulence similar to that which prevails in surface seawater, because such similarity suggests that the droplet-formation mechanisms will also be similar. Therefore, effectiveness testing should include measurement of droplet-size distributions, preferably in the presence of turbulent mixing energy, so that the observed size distribution will not be affected by size fractionation. Although droplet-size distributions have been measured in some lab-scale effectiveness-testing systems (Byford et al., 1984; Daling et al., 1990a; Lunel, 1995b; Fingas et al., 1995d), the effects of energy dissipation rate, oil and dispersant characteristics, and dispersant treatment should be more thoroughly investigated, because the existing database is not sufficient to support general conclusions regarding how (or whether) these factors affect the droplet-formation mechanisms and kinetics. Even fewer data are available regarding droplet-size distributions formed during dispersant effectiveness tests in wave tanks (Lunel, 1995b). Since one argument for increased use of these systems is their presumed ability to simulate sea surface conditions, it would be prudent to test this hypothesis by measuring droplet-size distributions and comparing them to those measured at sea. More effort has been focused on studying the relationship between droplet size and dispersant effectiveness, but conflicting results have been obtained. For example, one study demonstrated an inverse relationship between dispersant effectiveness and the volume median droplet diameter (Byford et al., 1984), whereas others observed no correlation between effectiveness and characteristic droplet size (Daling et al., 1990a; Fingas et al., 1995d; Lunel, 1995b). Although the relationship between effectiveness and droplet-size distribution is uncertain, the droplet-size distributions clearly vary among different experimental systems: volume mean diameters of about 3 mm were observed in a system that was mixed by a six-blade vaned-disk turbine (Jasper et al., 1978), whereas significantly larger diameters (volume median diameters of 20 to 45 µm) were observed in

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects FIGURE 3-3 Effect of a chemical dispersant (premixed Dasic Slickgone NS) on the droplet-size distribution produced when crude oil (Forties) was dispersed at sea. SOURCE: Reconstructed from data presented by Lunel (1995b).

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects experimental apparatuses that are more commonly used in dispersant testing (e.g., the Warren Springs Laboratory, Mackay-Nadeau-Steelman, and swirling flask tests) (Daling et al., 1990a; Fingas et al., 1995d). The strong dependence of droplet-size distributions on the characteristics of the experimental system are consistent with the hypothesis that they reflect microscale turbulence (Delvigne and Sweeney, 1988; Lunel, 1995b; Li and Garrett, 1998), and Lunel (1995b) suggested that laboratory-scale or wave-tank effectiveness tests should be evaluated based on their ability to produce size distributions similar to those observed at sea. In the ocean, turbulent energy is provided mainly by the wind, either by its direct action in shearing the water surface, or through the generation of surface waves. Above a critical wind speed, waves break, creating local areas of intense mixing. Internal waves, bottom shear stress caused by tidal or wind-driven currents interacting with a fixed bottom, and river inflows may also provide turbulent energy. Because of the variety of energy sources and mechanisms for oil droplet generation, it is unlikely that any single parameter can completely characterize the mixing energy responsible for oil dispersion. This is particularly true when including consideration of bench-scale lab tests (see below) in which mixing is produced by other mechanical means such as stirring, swirling, or tumbling. Nonetheless, the parameter that is most likely to be correlated with effective entrainment and dispersion is energy dissipation rate. Turbulent energy enters a water body at large length scales and is transferred to smaller scales by the process of vortex stretching until it is dissipated by viscosity into thermal energy at the smallest scales. At equilibrium, the rate of energy input equals the rate of energy transferred at each scale, and hence the rate of energy dissipation (Tennekes and Lumley, 1972). Energy dissipation rates can be expressed in units of energy loss per volume per time, e (J-m−3-s−1) where J is joules (kg-m2-s−2). So, the volumetric energy dissipation rate, e, can also be expressed as kg-m−1-s−3. The energy dissipation rate can also be expressed as energy loss per unit mass per time, denoted by ε (J-kg−1-s−1 or m2-s−3). The latter is numerically smaller than e by a factor of the water density (about 103 kg-m−3). Table 3-1, adapted from Delvigne and Sweeney (1988), gives approximate ranges of e and ε for a variety of water bodies. In-situ values of the dissipation rate can be determined from highly resolved velocity measurements. Doron et al. (2001) describe several methods involving either evaluation of fine-scale velocity gradients or finding a fit to the spectrum of turbulent kinetic energy (2) where E(k) is the turbulent kinetic energy density as a function of wave

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects treated slicks were generally 1–10 ppm and uniformly mixed down to 5 m (the maximum depth of measurement) (Lunel et al., 1997a). Within 5 km downcurrent of the grounding site, oil concentrations in the water column were 0.5–0.6 ppm throughout 4 days after termination of dispersant applications; by 12 days after termination, oil concentration were 0.2 ppm or lower. At distances of >10 km downcurrent, oil concentrations were 0.2–1.0 ppm throughout 6 days after termination of dispersant application; by 12 days after termination, they were 0.2 ppm or lower. There were reported mortalities of shallow sub-tidal and intertidal organisms, with bivalves and urchins washing up by the hundreds in some areas. Wild salmon, other finfish, crab, lobster, and whelk were found to have low levels of PAH but no taint. Intertidal mussels remained contaminated in one bay with heavy shoreline oiling for 19 months after the spill. SOURCE: Summarized from Harris (1997), Law et al. (1997), and Lunel et al. (1997a). data can be compared to values typically used in water accommodated fractions (WAF) generated for dispersed oil toxicity evaluations (see Chapter 5). Monitoring Dispersant Use During Actual Spills Monitoring of dispersant use means different things to different people. The mental model one has of concepts or definitions is generally associated with their background and stakeholder role. Dispersant-use monitoring can be separated into two basic categories: (1) information collected to help make timely operational decisions; and (2) data gathered for future analyses of fate and effect (Pond et al., 1997). Operational monitoring should provide information on the application platform’s spraying parameters and on whether or not oil is being entrained into the water column. This information should be conveyed immediately to those making the decision on whether or not to continue the operation. The second type of monitoring involves collecting data that can be later used to address the fate and effects of the dispersed oil and may also be used to ground truth some of the operational monitoring information (Hillman et al., 1997). In every dispersant application, operational monitoring is done to some degree. Depending on the circumstances, ground-truth information on fate and effect may or may not be required. Dispersant effectiveness is a phrase that has been interchangeably used to describe how well the product performs both in the laboratory

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects and in field applications. As discussed previously, there are three components that will determine dispersant effectiveness during spill response: operational effectiveness, chemical effectiveness, and hydrodynamic effectiveness. The common usage of “dispersant effectiveness” to describe performance in the laboratory and the field is unfortunate because laboratory-derived effectiveness usually does not equate to effectiveness in field applications (e.g., see Table 3-3). This dual usage has fostered misconceptions and misunderstanding throughout the response community and the public. As described previously in this chapter, laboratory tests generally measure chemical effectiveness, whereas effectiveness in the field is also dependent on operational and hydrodynamic factors. Therefore, a laboratory effectiveness of 60 percent does not mean that 60 percent effectiveness will be obtained in field applications. Depending on many factors, the field effectiveness for a product may range from 0 percent to 100 percent. Effectiveness of a dispersant application in the field has been defined as “the amount of the oil that the dispersant puts into the water column compared to the amount of oil that remains on the surface” considering the total amount of the oil that was treated (Fingas, 2002a,b; 2003; Lewis, 2004). U.S. Coast Guard, et al. (2001) define effectiveness based upon the amount of oil that the dispersant puts in the water compared to the amount of oil that was in the area treated. NRC (1989) concluded that a mass balance approach has given good effectiveness estimates in a few elaborate field tests, but “it is complicated, requires set-up time, and is not practical in real spills.” In field experiments, the release volume is known, the area of the slick can be measured, and the average thickness for this finite area can be calculated. In addition, dispersants are generally applied to the entire test slick; thus mass balance effectiveness estimates may be applicable. In accidental spills, however, only a portion of the total amount of spilled oil is normally treated, the oil thickness of the treated area is unknown and highly spatially variable, and thus the volume of oil in the treatment area is seldom known to any great accuracy in a timely manner. Presently, there is no valid and reliable method of determining slick thickness in the field, and any estimated value may easily be in error by an order of magnitude (Fingas, 2002a,b). Prior to development and implementation of a monitoring plan, it is imperative that the stakeholders agree on attainable goals and objectives for the monitoring (U.S. Coast Guard et al., 2001). Among these goals and objectives should be a working definition of field dispersant effectiveness and a set of Standard Operating Procedures (SOPs) with data quality objectives. The definition of field effectiveness could parallel the definitions of mechanical recovery (i.e., percent recovery of the entire spill) or in-situ burning (i.e., percent of oil burned from a contained area). The degree and extent of monitoring should be in proportion to the

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects sensitivity of the environment. In general, the more sensitive the environment, the more emphasis should be placed on monitoring. Sensitivity can be assigned based upon environmental and political parameters. Basically the resource trustees and stakeholders want to know how well the response works and the extent of the effects. There is a heightened concern as the sensitivity increases. Factors that have a direct relationship to sensitivity include, among others, nearness to shore, special habitats such as marine sanctuaries and parks, biological and migratory seasonality, size of incident, and nature of the spilled product. Nearness to shore generally involves environments with shallower water, lower dilution rates, higher productivity, fish and shellfish nursery grounds, higher concentrations of wildlife, greater commercial and recreational use, and shorter response times. As discussed in Chapter 2, it is very advantageous for the resource trustees and stakeholders to pre-identify sensitive areas, determine where and when dispersant use should be discussed, and outline monitoring objectives. Unless otherwise stated, pre-approval agreements generally are based on the assumption that use of dispersants, under specified conditions, will protect sensitive shoreline and water-surface resources without causing significant impacts to water-column and benthic resources, even assuming 100 percent dispersion of the slick. Operational Monitoring The primary reason to monitor operational aspects of dispersant use is to determine if the dispersant application is operationally effective (e.g., that the dispersant is being applied to the surface oil targets). The secondary purpose is to estimate the relative effectiveness of the operation (Fingas, 2003). Additional data also are needed to provide documentation on what dispersant was used, how much was used, when and where it was used, and the environmental conditions at the dispersal sites. Because there is no truly quantitative method to determine dispersant effectiveness in the field, the best that can be done is to qualitatively estimate if the dispersants are working (Henry, 2004). Effective/Ineffective Dispersant Applications It is assumed that some portion of the dispersant spray will miss the target due to wind drift of the spray or turning pumps on too soon or off too late (see earlier discussion of dispersant use in response to the T/V Exxon Valdez spill). Missing the target excessively should be documented in the monitoring report, and controls should be enacted to minimize this to within acceptable limits. An experienced trained observer is the best

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects TABLE 3-5 Guidelines to Assist in Determination of Effective/Ineffective Application Possible Dispersant Action Possible False Positives Possible False Negatives Difference in appearance between treated and untreated slick.a,b Suspended solids or algal blooms may resemble dispersed oil.a,b Dispersion may not be instantaneous, may take several minutes to a few hours to show dispersed plume.a,b,c,d Appearance of plume can range from brown to pale yellow.a,b,c,d Boat wakes through oil may appear as dispersed paths.a Visible cloud or plume not observed, water may be naturally murky.a,b,c Changes in area and thickness of the oil.c Dispersants may have a herding effect on thin oil. May also be seen as lacing.b,d Oil may be dispersing under the slick and not seen.b,d Higher fluorometer readings of dispersed oil in application area vs. background or non-treated slick area.a,b,c,d Rapidly dissipating whitish plume may be caused by dispersant alone (missed target).d     After initial visual assessment, some dispersed oil may resurface.d aUSCG et al., 2001. bNOAA, 1999. cExxonMobil, 2000. dFingas, 2003. way to assess if the dispersant operation is effective or not (Lewis and Aurand, 1997; U.S. Coast Guard et al., 2001; Fingas, 2003; Goodman, 2003; Henry, 2004). Even though there are difficulties with the interpretation of fluorometer data (Fingas, 2003; Goodman, 2003), the addition of confirmation fluorometer readings will help substantiate visual observations that there has been an increase in the amount of oil entrained into the water column under treated slicks. Table 3-5 contains guidelines to assist in determination of effective/ineffective application. Special Monitoring of Applied Response Technologies The protocol used by most if not all U.S. regions for obtaining operational monitoring information for dispersant use and in-situ burning is Special Monitoring of Applied Response Technologies (SMART) (U.S.

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects Coast Guard et al., 2001). The purpose of the dispersant section of SMART is to outline a protocol that rapidly can collect information to assist in real-time decisionmaking during dispersant applications (Barnea and Laferriere, 1999). SMART only outlines how to determine if the dispersant application is working, but provides no guidance on how to determine a percent dispersant effectiveness. It relies heavily on personnel being trained using job-aids developed to support SMART (Levine, 1999). For much of coastal and offshore waters of the United States, the resource trustees and stakeholders have designated selected areas as pre-approved for dispersant use. All pre-approved areas have a stipulation that requires use of the SMART protocols for operational monitoring, if operationally feasible. In an effort to better document effectiveness, field portable equipment has now been prepared and staged within various RRTs for immediate deployment in the event of a spill (Gugg et al., 1999; Barnea and Laferriere, 1999; Henry et al., 1999; Henry and Roberts, 2001). Some pre-approvals indicate that SMART will be used for fate and effects monitoring; however, SMART specifically states it “does not monitor the fate, effects, or impacts of dispersed oil” (U.S. Coast Guard et al., 2001). The SMART protocol contains three tiers of monitoring: Tier I is visual monitoring by a trained observer, preferably using an aircraft separate from the “spotter” aircraft directing the dispersant application (U.S. Coast Guard et al., 2001). The protocol recommends documentation via forms, photography, and videotape. Tier I monitoring may be enhanced through the use of remote sensing instruments, such as infrared thermal imaging, if data are available in real-time. The purpose of Tier I is to visually assess if the operation is working and rapidly report the findings to the decisionmakers. Typical observations include: (1) that the dispersant spray hit the slick; (2) a reduction in the amount of oil on the water surface after dispersant treatment; (3) a change in the appearance of the treated slick; and (4) the presence of a milky or cloudy plume in the water column. Tier II includes Tier I monitoring and adds an on-water component. From a vessel, water samples are analyzed via continuous flow fluorometer collecting water at a 1 m sampling depth. The protocol recommends comparing fluorometer measurements from three general areas: (1) background water outside the spill area, (2) below the surface oil slick before dispersant application, and (3) an area where the oil slick has been treated with dispersants. The purpose of Tier II is to confirm whether or not oil is being entrained into the water column (Barnea and Laferriere, 1999). A few water samples are collected for later laboratory analysis to validate and possibly quantify the fluorometer measurements. Tier III is presented as “Additional Monitoring” to collect information on transport and dispersion of the oil into the water column. It fol-

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects lows Tier II procedures but adds multiple depth fluorometer sampling of selected transects and provides for collection of additional environmental parameters, such as water temperature, conductivity, dissolved oxygen, pH, and turbidity. The SMART protocol includes collection of water samples to validate and quantify the fluorometer readings. Calibration methods and techniques are discussed in Lambert et al. (2001a,b) and Fingas (2002a,b). The validation method can estimate the quantity of “oil” in the water column, but the data cannot be used to differentiate between that part that is dissolved and that part that is in droplets. Fingas (2003, 2004a) discussed the precautions and proper use of fluorometry in the field. His comments on field techniques include awareness of possible contamination using Tygon tubing and maintaining the sampling probe in waters undisturbed by the vessel (in front of or outside the bow wave). The Alyeska Ship Escort Response Vessel System (SERVS) has developed dispersion monitoring guidelines that are similar to the SMART protocol, but the primary goal of the Alyeska/SERVS protocol is to provide real-time assessment of the environmental effects of dispersion (Hillman et al., 1997). The Alyseka/SERVS protocol relies on aerial monitoring as the primary tool for monitoring dispersant effectiveness and effects with additional support provided by collection of water samples and in-situ fluorometry. This protocol is not intended to provide quantitative estimates of dispersant effectiveness, real-time estimates of water-column dispersed oil concentrations, or estimates of oil mass balance. This protocol attempts to monitor the dispersed oil plume by locating the water-column sampling stations and the in-situ fluorometry transect relative to drogues that drift with subsurface currents (usually at 2-m depth). Whereas SMART and the Alyeska/SERVS protocols rely on conventional filter fluorometers with a single filter for excitation and another for emission for in-situ measurement of dispersed oil concentrations, multiple-wavelength fluorometers and in-situ instruments capable of measuring particle-size distributions have been investigated for research use (Fuller et al., 2003). Unfortunately, the performance of these instruments for monitoring oil dispersion at sea has not yet been evaluated (Ojo et al., 2003). Additional Operational Monitoring To better document the operation and to possibly provide clues to future questions, several delivery platform and environmental parameters should be recorded. Pre-application documentation should include the name, lot number, and quantity of dispersant loaded on the aircraft or vessel. A sample should be taken, with proper chain-of-custody, from each dispersant lot (to allow for later analysis if verification of product effec-

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects tiveness is needed). After each sortie, the amount of dispersant remaining onboard should be documented. Also, other data are needed on the platform performance during the application and on the environmental conditions in the application area. Table 3-6 provides guidance on the additional monitoring data or samples to be obtained. Most of the performance data should be automatically recorded on the platform. Environmental Monitoring As discussed in Chapter 2, there is a reasonable degree of confidence in the current ability to assess trade-offs, relative to use of dispersants, in offshore waters. In general, offshore waters are considered to be less sensitive to dispersed oil impacts, because of rapid dilution of dispersed oil, than shallower or nearshore environments. But as shallower or nearer to shore waters are evaluated for dispersant use, the sensitivity of the environment and the degree of uncertainty make the assessment more difficult. The database of oil component acute toxicity is much better than the knowledge of the bioavailability of dispersed oil components in the water column. Unfortunately, most of the measurements on concentrations of TABLE 3-6 Guidance on the Additional Monitoring Data or Samples Pre-application Application Post-application Name of dispersant Dispersant lot number Sample of each dispersant lot number Volume of dispersant onboard Platform (aircraft or vessel) Description of spray system Dispersant pump calibration documentation Spray nozzle test documentation Spray time Pump rate Speed during application Spray height during application Sample of weathered and neat oil Dispersant applied neat or diluted Wind speed and direction Current speed and direction Air and surface water temperature Cloud cover Surface salinity Wildlife in area (birds, mammals, turtles) Approximate spray width Approximate spray path length Number of passes over same area to achieve adequate dispersion. Sea state Volume of dispersant remaining onboard

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects dispersed oil in open water are from fluorometer readings or from the total extraction of unfiltered water samples. Thus at best, the total concentration of oil components in the water column is known, but not whether the component concentrations reside in the water or in oil droplets is not known. Questions the risk assessors need answers to concerning the dispersed oil include: (1) How are the components of dispersed oil distributed in the water column? and (2) What fractions are in the dissolved phase and what fractions are in droplets or adhered to particulates in the water column? These data cannot be obtained through fluorometry, and Page et al. (2000b) have shown that estimations of oil-component partitioning based upon solubility coefficients alone are not reliable for oil-in-water mixtures. The data can be obtained via discrete large-volume water samples that are collected and filtered immediately to differentiate between components that are truly dissolved and those that are present as dispersed oil droplets (Payne et al. 1999; Payne and Driskell, 2001, 2003). These samples, at a minimum, should be analyzed for dispersed oil droplet and dissolved-phase PAH and total petroleum hydrocarbon (TPH) concentrations in the filtered and unfiltered water. Sample collection should be from several depths and repeated over time. Real time in-situ fluorometry data should be used to locate where to take samples and to verify that the discrete samples were taken in the dispersed oil plume. In additional to finite grab examples collected with traditional water-sampling equipment, aliquots of effluent from the fluorometers should also be collected for chemical analysis. Whenever possible, separate fractions for dissolved and particulate/oil-phase components should also be analyzed (Payne et al., 1999; Payne and Driskell, 2001, 2003). Monitoring data, coupled with local transport mechanisms, can be used to validate computer-model predictions, and thus reduce the uncertainty of the fate of dispersed oil components. Ultimately, this will provide decisionmakers with a better tool to assess use of dispersants in sensitive environments. The trustees of the local resources at risk will determine if other types of monitoring are needed to assess the effects. The extent of monitoring should be based on the sensitivity of the environment and the predicted amount of dispersed oil reaching the resources of concern. The collection and analysis of samples, whether they are sediment, nekton, or benthos, should be conducted so there can be a direct comparison with water-column analytes. DEVELOPING ADEQUATE UNDERSTANDING OF DISPERSANT EFFECTIVENESS TO SUPPORT DECISIONMAKING As discussed in Chapter 2 and shown in Figure 2-4, the potential effectiveness of dispersants is a key consideration at several steps in the

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects decision-making process. Significant work has been done to test dispersant products on a range of oil types or refined products under different test conditions (temperature, salinity, etc.). The test protocols were designed to establish a high degree of reproducibility, but were never intended to replicate actual environmental conditions that may be encountered during a spill. However, these kinds of tests are useful to provide guidance on whether or not a test oil is likely to be dispersible under ideal conditions. The fourth question in Figure 2-4—“Are conditions conducive?”—addresses the range of factors that affect the overall field effectiveness of dispersant application once the oil starts to spread and weather. Currently, predicted dispersant effectiveness for a specific spill event is based on simple models and past experience. In current fate and transport models, dispersant effectiveness is an input value. In the future, it would be desirable to possess the ability to predict dispersant effectiveness over time through the use of a physical-chemical efficiency model. However, additional research is needed to develop the model. Relevant state and federal agencies and industry should develop and implement a focused series of studies that will enable the technical support staff advising decisionmakers to better predict the effectiveness of dispersant application for different oil types and environmental conditions over time. Bench-scale effectiveness tests can provide a valuable tool for investigating the factors and interactions that affect the chemical effectiveness of oil dispersion. A particular strength is their ability to inexpensively and quickly test a large number of conditions. Currently, most bench-scale effectiveness tests incompletely characterize the test conditions and do not systematically vary factors, such as mixing energy, that are known to have a strong influence on the process of oil dispersion. In addition, important response variables, such as oil droplet-size distributions, are not routinely measured. As a result, bench-scale effectiveness tests cannot, in general, provide the type of input that is needed for fate and transport models. Experimental systems used for bench-scale effectiveness tests should be characterized to determine the energy dissipation rates that prevail over a wide range of operating conditions. Future effectiveness tests should measure chemical effectiveness over a range of energy dissipation rates to characterize the functional relationship between these variables. Finally, evaluation of chemical effectiveness should always include measurement of the droplet-size distribution of the dispersed oil. Wave-tank-scale effectiveness tests are particularly useful for investigating factors that cannot be studied in laboratory-scale tests. In addition, the more realistic mechanism of energy input in experiments conducted in wave tanks reduces the sensitivity of results to uncertainties

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects regarding the mechanism of oil-droplet formation and, therefore, scaling of laboratory- or wave-tank-derived effectiveness estimates to sea-surface conditions. The design of wave-tank dispersant-effectiveness studies should specifically test hypotheses regarding factors that can affect operational effectiveness. These factors include oil properties that are representative of those expected to prevail under spill-response conditions, such as water-in-oil emulsification and the potential for heterogeneity in the rheological properties of the floating oil (e.g., formation of a “skin” that resists dispersant penetration). Dispersant droplet-size distributions and impact velocities should be similar to those that would be expected to be generated by dispersant application methods commonly used in oil-spill response. Tank tests that determine the ability of mechanical recovery methods to recover oil that has been treated with dispersant but not effectively dispersed, or re-floated oil, should be carried out. A more complete understanding of what limitations the unsuccessful use of dispersants may have on subsequent mechanical recovery methods could greatly reduce concern over relying on operational testing of the dispersant effectiveness in the early phases of spill response. Energy-dissipation rates should be determined for wave tanks over the range of operating conditions that will be used in dispersant effectiveness tests. The wave conditions used in dispersant effectiveness tests should represent a specific environment of interest. It may be necessary to conduct experiments over a range of energy dissipation rates to adequately represent the environment of interest. More robust understanding of dispersant effectiveness can be derived from test tanks, if more rigorous protocols are implemented that better quantify the eventual fate of the test oil. The concentration of oil should be measured in all identifiable compartments to which it could be transferred when dispersant effectiveness is investigated in wave tanks. This includes, but may not be limited to, the water surface, the water column, the atmosphere, and wave-tank surfaces. Oil mass balances should be reported in an effort to better understand the accuracy of effectiveness quantification. In addition, the droplet-size distribution of the dispersed oil should be measured and reported. Little is known of the potential leaching of surfactant from floating oil and dispersed oil droplets at realistic oil-to-water ratios and under turbulence conditions that might be encountered in the field. In particular, the effects of surfactant leaching on the effectiveness of oil dispersion and the potential for droplet coalescence should be understood better. Coalescence and resurfacing of dispersed oil droplets as a function of mixing time should be studied in flumes or wave tanks with high water-to-oil

OCR for page 51
Oil Spill Dispersants: Efficacy and Effects ratios (to promote leaching of surfactant into the water column). Periods of wave-induced turbulence should be followed by periods of relative calm to allow droplets to resurface. The surfactant concentration remaining in the resurfaced oil should be measured, and its dispersibility should be measured (by introducing more wave turbulence) to evaluate the ultimate fate of resurfaced oil. Alternatively, oil dispersion should be measured after dispersant is applied and incubated with floating oil under calm conditions to determine the effect of surfactant leaching from a surface oil film on dispersant effectiveness. Although careful and controlled research in the laboratory or test tank will be important to developing tools to support decisionmaking, the results of dispersant application during real spills will be the most important indicator of whether or not the dispersant application was effective. Field data are essential to a better understanding of the spill-specific conditions that affected the dispersant operation, and they should be used to validate model predictions. To improve the quality of field data collected during dispersant applications, more robust monitoring capabilities should be implemented. Specific attention should be given to: Developing an environmental monitoring guidance manual for dispersant application monitoring with suggested sampling and analytical techniques, sampling methods, and QA/QC to ensure cost effectiveness and maximum utilization of the data Developing a detailed standard operating procedure (including instrument calibrations and data quality objectives) for each sampling and analytical module (SMART is guidance only) Developing a definition of field effectiveness Measuring dispersed oil droplet and dissolved-phase TPH and PAH concentrations with grab samples of filtered and unfiltered water (these data can then be compared to model predictions and toxicity data for both dissolved and particulate/oil-phase components) as a function of location and time.